Split flow issue with GVF-Rational solver

Hello Jared,

Split flow with diversion rating is still possible with the GVF-Rational solver. You may have heard about it in this article:

Modeling a flow split (diversion) with the SewerCAD or StormCAD numerical solver

See the "caveats" in the above article for possible reasons why it may not be working for you. Additionally you can read about how the Implicit and Explicit dynamic solvers available in SewerGEMS will handle split flow automatically, in which case any diversion rating curves will indeed be ignore (so you may want to confirm which solvers you are using, in the calculation options.) If this does not help, there may be some other problem with the configuration - in this case, please provide the exact message you are encountering, and a copy of the model plus steps to reproduce, if possible: Sharing Hydraulic Model Files on the OpenFlows Forum

Note that the latest version of SewerGEMS is 10.02.03.03. I am not aware of any changes related to diversion handling with the StormCAD solver in the CONNECT Edition release. See more about upgrading here: Downloading OpenFlows Software

Hi Jesse, 

I just uploaded the model files for your review. 

The errors that we were receiving were on link "CO-1695_E3" but I don't seem to be getting the errors now, so it might have been my colleagues computer. The errors said something like "Control structure in diversion link not supported" and that it was being ignored. I'm very curious to why it has now changed, but the pipe properties window, under "Diversion Type", still has a description that says "The type of diversion to use. Note that only the Rating Table type is supposed by the Dynamic Wave (implicit) engine. The SWMM engine supports all Diversion Types.". This led us to believe that the GVF-Rational solver does not support the diversion link - hence the error that we were seeing. 

I would still request that you review the model and see if you see any issues with the solver and the diversion link. We have an outfall structure directly downstream of the diversion link, and even though the model now shows a diverted flow in the CO-1695_E3 link, the outfall does not show any flow which doesn't seem to make sense. 

Thanks, 

Jared

Hi Jared,

I too do not see any user notification in the model related to that conduit. 

The message you mentioned would seem to indicate that someone attempted to set a conduit control structure (like a weir or orifice - this is separate from the diversion rating table) on the same conduit as the diversion rating table, then attempted to use one of the solvers that support diversion rating tables.

With the GVF-Rational solver though (which is what the model was set to use), conduit control structures are not supported (see note at bottom of above article), so the input field does not appear. Even if you change the solver to Implicit for example, the start and stop conduit control structure fields are set to false. Maybe the colleague that had the issue had changed the solver or conduit control structure configuration, or was using a different version?

The property grid description you mentioned for the "diversion type" is trying to say that when using the Implicit solver, you can only select the rating table type. The GVF-Rational solver supports it as well, plus other diversion types. 

I do see one issue with that particular conduit though - in the "POST-RATIONAL" scenario, the flow approaching the diversion (79.85 CFS) is less than the lowest value in the rating table (80.00 CFS), so it actually results in a small negative flow when interpolated. You'll want to add a 0,0 point at the top of the rating table to avoid this. You may also want to consider adding more flow points on the rating table to cover the maximum possible upstream flow. 

Note that I used the latest version (10.02.03.03) when reviewing the model, whereas you appear to be on an older version (10.01.01.04). I see you mentioned that you are seeing "diverted flow" in that conduit, so if you were not seeing the slightly negative flow that I was seeing (explanation above), then there may be differences in the version, or perhaps the wrong model version was sent?

Thanks for taking a look. The colleague of mine is unfortunately out of office for a week so I cannot confirm what happened but will try and follow up. 

After reading your post, I believe what may have happened is an old child scenario that we had setup used the implicit solver and likely had the conduit set as a control structure.. and I'm wondering if when we deleted that scenario, the conduit kept it's "control structure" property in some kind of way which gave us the error. And since you explained that the input field does not appear within the GVF-Rational solver, we were unable to see that property or remove it.

What I think may have solved the issue that I want to confirm when my colleague is back in the office is he may have deleted the conduit link and re-added a new conduit link which would have reset the properties and therefore not had this hidden "control structure" property which we were unable to see and remove. 

Thanks for the note on the rating table as well. I'll make sure that is revised as well. I had noticed the reverse flow arrow as well and played around with the table before sending my follow up response, hence the diverted flow that I was seeing. 

One more follow up - is there a way to get the outfall structure to show diverted flow? I think I saw a property description for the outfall structure's flow value that says it does not include diverted or overflow which likely answers my question, but ideally we would want to see that diverted flow in our Outfall flex table.. 

Jared Sandberg said:

is there a way to get the outfall structure to show diverted flow?

In the past, you could work around this by inserting a transition element to introduce a second conduit between the diversion conduit and the outfall (splitting the conduit with the transition element, then unselecting the diversion option in the second conduit), but I see this does not seem to work in the latest releases. I have added a note to split flow wiki article for now, but will check with my colleagues to confirm if this is working as-designed. 

I tried a few other things but was not able to find another workaround to include the diverted flow in the outfall node's calculated flow field. For now, you will need to look at the flow in the diversion conduit.

Jared Sandberg said:

is there a way to get the outfall structure to show diverted flow?

In the past, you could work around this by inserting a transition element to introduce a second conduit between the diversion conduit and the outfall (splitting the conduit with the transition element, then unselecting the diversion option in the second conduit), but I see this does not seem to work in the latest releases. I have added a note to split flow wiki article for now, but will check with my colleagues to confirm if this is working as-designed. 

I tried a few other things but was not able to find another workaround to include the diverted flow in the outfall node's calculated flow field. For now, you will need to look at the flow in the diversion conduit.